Always Knowing What to Say Next
Get the latest in your inbox
As an AI startup founder and CEO, I look for simple, but still very true, images or metaphors or analogies to help business leaders understand how the right tech strategy can impact their business.
When it comes to Generative AI and enterprise software, this can be challenging, not least since the tech is frothy dynamic and the incumbents are aggressive and the space is very noisy. But I also believe, like a good cartesian rationalist, that the truth is always clear and simple; or that it always can be.
What’s the simplest true thing that every business leader needs to know about Generative AI? What if there were a machine that always knew what to say next? And what if, in a great many cases, knowing what to say next means this machine knows what to do next?
That’s what Generative AI and LLMs are in a nutshell:
LLMs are machines for telling people (and other machines) what to say next and in some cases what to do next, given what’s already been said and what’s already been done.
And in fact this explains a great deal of the work to be done to profit from AI in the enterprise: how do we arrange the data and business process landscape such that LLMs are part of the ongoing streams of conversation, both internally and externally, so that they can tell us what to say (and sometimes do) next?
In other words, what to say next is a function of what has been said so far; that’s exactly (and exclusively) how AI hinges on your enterprise data.
Even more crucially for business leaders: which are the parts of your business where knowing what to say or do next means that you win? Customer service? Sales? Design? Financial risk management and compliance? Quarantined batch management in pharma? Complex product configuration in automotive supply chain?
Everything else is for the engineers to figure out! The winners of the AI era on the buy side will be the organizations that first internalize this simple but profound change in the world and second, operationalize it.
